Package cedarview-drm

Bug #1025720 reported by Jani Monoses
38
This bug affects 5 people
Affects Status Importance Assigned to Milestone
cedarview-drm-drivers (Ubuntu)
Fix Released
Undecided
Alberto Milone

Bug Description

Intel Cedarview based platforms need closed source binary drivers for accelerated graphics support.
There are 3 packages to be added to 12.04 to support laptops using such hardware.

Description: Cedar Trail drm driver in DKMS format.
 This package contains the drm driver for Cedar Trail in DKMS format.

Revision history for this message
Ubuntu Foundations Team Bug Bot (crichton) wrote :

Thank you for taking the time to report this bug and helping to make Ubuntu better. It seems that your bug report is not filed about a specific source package though, rather it is just filed against Ubuntu in general. It is important that bug reports be filed about source packages so that people interested in the package can find the bugs about it. You can find some hints about determining what package your bug might be about at https://wiki.ubuntu.com/Bugs/FindRightPackage. You might also ask for help in the #ubuntu-bugs irc channel on Freenode.

To change the source package that this bug is filed about visit https://bugs.launchpad.net/ubuntu/+bug/1025720/+editstatus and add the package name in the text box next to the word Package.

[This is an automated message. I apologize if it reached you inappropriately; please just reply to this message indicating so.]

tags: added: bot-comment
Changed in ubuntu:
assignee: nobody → Alberto Milone (albertomilone)
Revision history for this message
Steve Langasek (vorlon) wrote :
Download full text (4.2 KiB)

Hi Alberto,

Reviewing the package in the new queue, I spotted a few issues. The first few of these need to be fixed before I will accept the package into precise. Can you please take care of these and reupload?

Depends: [...] findutils (>= 4),

findutils is an Essential package, and version 4 is older than hardy. This dependency should be dropped.

debian/grub.d/11_custom_cmdline: this is NIHing the grub-gfxpayload blacklist handling already implemented in grub via /usr/share/grub-gfxpayload-lists/blacklist. Instead of shipping a file under /etc/grub.d which hard-codes disabling of vt handoff even when the hardware is not present, please add a blacklist file to this directory like the ones in the grub-gfxpayload-lists package and call 'update-grub-gfxpayload' (instead of update-grub') from your package's postinst and postrm (on remove/purge only, and guarded with || true). Creating the blacklist should be straightforward, you're already processing a list of PCI IDs in debian/rules.

You should add a dependency on grub-gfxpayload-lists for this.

debian/lightdm.conf: why is this necessary? including this file in the package means that merely installing the driver package causes unity3d to stop working out of the box, even if this video hardware is not present. (The same reason not to hard-disable vt.handoff.) Does the Unity auto-detection of 3D support not work correctly on this hardware for some reason? If it's *really* confirmed to be needed, then ok; but we should definitely use the autodetection if possible.

lintian reports the following issues with the maintainer scripts:

W: cedarview-drm: possible-bashism-in-maintainer-script postinst:94 '&>'
W: cedarview-drm: command-with-path-in-maintainer-script postrm:39 /usr/sbin/update-initramfs

Looking more closely, I see that this section of the postinst is entirely redundant with what dh_dkms is supposed to automatically add. But dh_dkms is currently a complete no-op, because it requires the conf file to be present as debian/dkms, not debian/dkms.conf. So at a minimum, the bashisms in the postinst need to be fixed - but it would be better to remove this section of the postinst, rename debian/dkms.conf.in to debian/dkms.conf, and leverage dh_dkms that you're already build-depending on and calling. This also removes the need for an auto-generated debian/prerm.

Below are some further suggestions on improving the readability / maintainability / correctness of the package. I don't consider these blockers for accepting the package into precise, but I would appreciate seeing them fixed in a subsequent upload to quantal.

%:
        dh --with dkms $@

Note the ordering of options to dh: in recent versions, the target $@ must come as the first argument.

%:
        dh $@ --with dkms

override_dh_auto_build:
clean: regen-from-templates
        dh_quilt_unpatch
        rm -f $(CURDIR)/debian/dkms.conf

        dh clean

I think this should be written:

override_dh_auto_build:
override_dh_auto_clean: regen-from-templates

- This is already a 3.0 (quilt) format source package, so any quilt handling here is redundant. (the build-dependency should also be dropped.)
- You don't need to rm -...

Read more...

tags: added: patch
Revision history for this message
Alberto Milone (albertomilone) wrote : Re: [Bug 1025720] Re: Package cedarview-drm

On 27/07/12 23:21, Steve Langasek wrote:
> Hi Alberto,
>
> Reviewing the package in the new queue, I spotted a few issues. The
> first few of these need to be fixed before I will accept the package
> into precise. Can you please take care of these and reupload?
>
> debian/grub.d/11_custom_cmdline: this is NIHing the grub-gfxpayload
> blacklist handling already implemented in grub via /usr/share/grub-
> gfxpayload-lists/blacklist. Instead of shipping a file under
> /etc/grub.d which hard-codes disabling of vt handoff even when the
> hardware is not present, please add a blacklist file to this directory
> like the ones in the grub-gfxpayload-lists package and call 'update-
> grub-gfxpayload' (instead of update-grub') from your package's postinst
> and postrm (on remove/purge only, and guarded with || true). Creating
> the blacklist should be straightforward, you're already processing a
> list of PCI IDs in debian/rules.
>
> You should add a dependency on grub-gfxpayload-lists for this.
>

The main reason why I didn't use this was that it gives us a black
screen until we get a 1 or 2 seconds plymouth screen and then we get to
the lightdm screen. At least this is what happens on our machines.

I'll accept your changes just in case other machines with supported
chipsets react in a
different way.

> debian/lightdm.conf: why is this necessary? including this file in the
> package means that merely installing the driver package causes unity3d
> to stop working out of the box, even if this video hardware is not
> present. (The same reason not to hard-disable vt.handoff.) Does the
> Unity auto-detection of 3D support not work correctly on this hardware
> for some reason? If it's *really* confirmed to be needed, then ok; but
> we should definitely use the autodetection if possible.

The driver's OpenGL implementation is, well, partial ;). This means that
Unity will start (as the driver passes the Unity's tests) but you won't
see anything useful on your screen.

> - debian/dirs is probably entirely pointless. You should only need this
> file if you are shipping empty directories in the package.

Good point

>
> - debian/copyright references http://dep.debian.net/deps/dep5/, which is
> an obsolete draft. This should ideally be updated to conform to
> copyright-format 1.0 (http://www.debian.org/doc/packaging-manuals
> /copyright-format/1.0/).

Ok, I'm attaching the new diff with your changes (thanks for all your
work, BTW) and mine.

Revision history for this message
Steve Langasek (vorlon) wrote :

On Tue, Jul 31, 2012 at 02:53:49PM -0000, Alberto Milone wrote:
> > debian/grub.d/11_custom_cmdline: this is NIHing the grub-gfxpayload
> > blacklist handling already implemented in grub via /usr/share/grub-
> > gfxpayload-lists/blacklist. Instead of shipping a file under
> > /etc/grub.d which hard-codes disabling of vt handoff even when the
> > hardware is not present, please add a blacklist file to this directory
> > like the ones in the grub-gfxpayload-lists package and call 'update-
> > grub-gfxpayload' (instead of update-grub') from your package's postinst
> > and postrm (on remove/purge only, and guarded with || true). Creating
> > the blacklist should be straightforward, you're already processing a
> > list of PCI IDs in debian/rules.

> > You should add a dependency on grub-gfxpayload-lists for this.

> The main reason why I didn't use this was that it gives us a black
> screen until we get a 1 or 2 seconds plymouth screen and then we get to
> the lightdm screen. At least this is what happens on our machines.

Ok. When you do it the other way, manually disabling vt.handoff but leaving
gfxmode=keep, what do you see instead / what do you expect to see? I guess
that you might see a *purple* screen for a few seconds, followed by a
flicker and plymouth, followed by lightdm. But if this difference between
purple and black is important, please file a bug against the grub2 package
to discuss. This combination of gfxmode=keep + no vt_handoff is not tested
in Ubuntu, and I don't think it should be introduced in a hardware-specific
way.

> > debian/lightdm.conf: why is this necessary? including this file in the
> > package means that merely installing the driver package causes unity3d
> > to stop working out of the box, even if this video hardware is not
> > present. (The same reason not to hard-disable vt.handoff.) Does the
> > Unity auto-detection of 3D support not work correctly on this hardware
> > for some reason? If it's *really* confirmed to be needed, then ok; but
> > we should definitely use the autodetection if possible.

> The driver's OpenGL implementation is, well, partial ;). This means that
> Unity will start (as the driver passes the Unity's tests) but you won't
> see anything useful on your screen.

Right, understood... :)

> > - debian/copyright references http://dep.debian.net/deps/dep5/, which is
> > an obsolete draft. This should ideally be updated to conform to
> > copyright-format 1.0 (http://www.debian.org/doc/packaging-manuals
> > /copyright-format/1.0/).

> Ok, I'm attaching the new diff with your changes (thanks for all your
> work, BTW) and mine.

Looks good to me. Please reupload.

Revision history for this message
Alberto Milone (albertomilone) wrote :

I've just re-uploaded the source with your changes and a new upstream release (which contains a minor fix). See the attached file for the complete diff.

Changed in ubuntu:
status: New → In Progress
Revision history for this message
Alan Searchwell (searchie) wrote :

In July Intel "released" graphics drivers for Cedar Trail. There is a file available on the Intel web site with the file name "cdv-gfx-drivers-1.0.1_bee.tar.bz2" and a date of July 6, 2012. It can be found by searching the Intel Download Center (http://downloadcenter.intel.com/Default.aspx?lang=eng) for the filename or the string "Linux* PowerVR Graphics/Media Drivers". The download page links to the file, release notes and a link, "Enabling hardware accelerated playback" that, takes one to a page containing links to two pdf documents titled "Enabling Hardware Accelerated Playback for Intel® AtomTM Processor N2000/D2000 Series", one for Ubuntu and one for Fedora.

Is this significant?

Revision history for this message
Jani Monoses (jani) wrote :

Alan, this is the very same binary packaged, with a diffrenent versioning due to historical reasons (Intel only recently started publicly versioning the blobs AFAIK)

Revision history for this message
Alusa (alusa) wrote :

We installed the cedarview-drm from the repos on our laptop and the computer hangs during boot, is this being fixed?

Revision history for this message
Adam Conrad (adconrad) wrote : Update Released

The verification of this Stable Release Update has completed successfully and the package has now been released to -updates. Subsequently, the Ubuntu Stable Release Updates Team is being unsubscribed and will not receive messages about this bug report. In the event that you encounter a regression using the package from -updates please report a new bug using ubuntu-bug and tag the bug report regression-update so we can easily find any regresssions.

Revision history for this message
Steve Langasek (vorlon) wrote :

I've had verbal confirmation from James Leddy that he's tested with the full set of packages in precise-proposed and the install test worked through jockey; the packages have been copied to precise-updates, so considering this verified and resolved.

Changed in ubuntu:
status: In Progress → Fix Released
tags: added: verification-done
Revision history for this message
Giuseppe Terrasi (giuseppeterrasi-deactivatedaccount) wrote :

Hi

I have switched from the ppa packages [1] to the repository version.

jockey-text -l
xorg:cedarview_gfx - Intel Cedarview graphics driver (Proprietario, Abilitato, In uso)
kmod:cedarview_gfx - drm driver for the Intel GMA500 (Proprietario, Abilitato, In uso)

For this graphic card: Intel Corporation Atom Processor D2xxx/N2xxx Integrated Graphics Controller (rev 09)

it work:
1) with no pae
2) no 3d effect
3) aixgl off

Is what you expected? You can confirm that 3D is not supported for this graphic card at this moment on 3.2.0 kernel?

[1] https://launchpad.net/~sarvatt/+archive/cedarview/

Revision history for this message
Mind Booster Noori (marado-isp) wrote :

Hi there,

It seems that the cedarview packages are only available for precise, but not for quantal. Is there any reason for that?

Revision history for this message
Jani Monoses (jani) wrote :

yes, they only work with the 3.2 kernel as that is what Intel tested the proprietary drivers with.
I don't know how much effort it would be to get them working with 3.5 or newer.

sruthinlal (sruthinlal)
affects: ubuntu → cedarview-drm-drivers (Ubuntu)
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.